顾及光谱特性的多光谱与全色影像HIS融合方法

A fusion method of multispectral images and panchromatic image based on HIS transform considering spectral characteristics

应用GHIS(Generalized-Hue-Intensity-Saturation)变换融合多光谱影像与全色影像中存在的光谱扭曲问题,主要是由于变换后的I分量与全色影像之间差异过大。为此,本文提出了一种基于统计关系定权的I分量计算方法,从数据整体角度考虑,在不能获取I分量与多光谱影像准确函数关系情况下,通过各波段影像与全色影像的相关系数定权计算I分量,使新的I分量与全色影像尽可能接近,然后通过GHIS变换完成融合。试验选取了两组不同传感器卫星影像进行验证,经视觉评价与统计分析,均表明本文方法相对于以往HIS改进方法,能够更好地保持原多光谱影像的特性,对全色影像的空间细节也能够较好集成,但与理想的真实多光谱影像相比,其光谱特征仍有一定的差距。

The reasons that causes spectral distortion in various HIS transform methods in the fusion of multi-spectral images and pan- chromatic image are the differences between I component and panchromatic image. A method was proposed to calculate I component using the weights decided by correlation coefficients between multi-spectral images and panchromatic image in the case of no exact functional relation, and made I component as closely to panchromatic image. Two different data sets were used to evaluate the proposed fusion method. The fusion results were evaluated visually and statistically. The results showed that the method proposed in this paper compared with the previous HIS method improved can better preserve the characteristics of original multi-spectral images. The spatial details of the panchromatic image could also be integrated very well. But the fusion results comparing with the ideal multi-spectral images still could lose some details of spectral features.